Tests: multiple fixes, increase timings

.travis.yml

@@ -29,10 +29,10 @@ jobs:
         - bash <(curl -s https://codecov.io/bash)
     - name: "Main Tests with raft consensus"
       script:
-        - go test -v -failfast -consensus raft .
+        - travis_wait go test -v -timeout 15m -failfast -consensus raft .
     - name: "Main Tests with stateless tracker"
       script:
-        - go test -v -failfast -tracker stateless .
+        - travis_wait go test -v -timeout 15m -failfast -tracker stateless .
     - name: "Golint and go vet"
       script:
         - go get -u golang.org/x/lint/golint

config_test.go

@@ -34,33 +34,33 @@ var testingClusterCfg = []byte(`{
     "state_sync_interval": "1m0s",
     "ipfs_sync_interval": "2m10s",
     "replication_factor": -1,
-    "monitor_ping_interval": "350ms",
-    "peer_watch_interval": "200ms",
+    "monitor_ping_interval": "1s",
+    "peer_watch_interval": "1s",
     "disable_repinning": false,
     "mdns_interval": "0s"
 }`)
 
 var testingRaftCfg = []byte(`{
     "data_folder": "raftFolderFromTests",
-    "wait_for_leader_timeout": "10s",
+    "wait_for_leader_timeout": "5s",
     "commit_retries": 2,
     "commit_retry_delay": "50ms",
     "backups_rotate": 2,
     "network_timeout": "5s",
-    "heartbeat_timeout": "200ms",
-    "election_timeout": "200ms",
-    "commit_timeout": "150ms",
+    "heartbeat_timeout": "700ms",
+    "election_timeout": "1s",
+    "commit_timeout": "250ms",
     "max_append_entries": 256,
     "trailing_logs": 10240,
     "snapshot_interval": "2m0s",
     "snapshot_threshold": 8192,
-    "leader_lease_timeout": "200ms"
+    "leader_lease_timeout": "500ms"
 }`)
 
 var testingCrdtCfg = []byte(`{
     "cluster_name": "crdt-test",
     "trusted_peers": ["*"],
-    "rebroadcast_interval": "150ms"
+    "rebroadcast_interval": "250ms"
 }`)
 
 var testingBadgerCfg = []byte(`{
@@ -114,12 +114,12 @@ var testingTrackerCfg = []byte(`
 `)
 
 var testingMonCfg = []byte(`{
-    "check_interval": "550ms",
+    "check_interval": "800ms",
     "failure_threshold": 6
 }`)
 
 var testingDiskInfCfg = []byte(`{
-    "metric_ttl": "350ms",
+    "metric_ttl": "900ms",
     "metric_type": "freespace"
 }`)
 

ipfscluster_test.go

@@ -58,7 +58,8 @@ var (
 	ptracker  = "map"
 	consensus = "crdt"
 
-	testsFolder = "clusterTestsFolder"
+	ttlDelayTime = 2 * time.Second // set on Main to diskInf.MetricTTL
+	testsFolder  = "clusterTestsFolder"
 
 	// When testing with fixed ports...
 	// clusterPort   = 10000
@@ -124,6 +125,10 @@ func TestMain(m *testing.M) {
 		}
 	}
 
+	diskInfCfg := &disk.Config{}
+	diskInfCfg.LoadJSON(testingDiskInfCfg)
+	ttlDelayTime = diskInfCfg.MetricTTL * 2
+
 	os.Exit(m.Run())
 }
 
@@ -298,14 +303,15 @@ func createHosts(t *testing.T, clusterSecret []byte, nClusters int) ([]host.Host
 	dhts := make([]*dht.IpfsDHT, nClusters, nClusters)
 
 	tcpaddr, _ := ma.NewMultiaddr("/ip4/127.0.0.1/tcp/0")
-	quicAddr, _ := ma.NewMultiaddr("/ip4/127.0.0.1/udp/0/quic")
+	// Disable quic as it is proving a bit unstable
+	//quicAddr, _ := ma.NewMultiaddr("/ip4/127.0.0.1/udp/0/quic")
 	for i := range hosts {
 		priv, _, err := crypto.GenerateKeyPair(crypto.RSA, 2048)
 		if err != nil {
 			t.Fatal(err)
 		}
 
-		h, p, d := createHost(t, priv, clusterSecret, []ma.Multiaddr{quicAddr, tcpaddr})
+		h, p, d := createHost(t, priv, clusterSecret, []ma.Multiaddr{tcpaddr})
 		hosts[i] = h
 		dhts[i] = d
 		pubsubs[i] = p
@@ -382,6 +388,7 @@ func createClusters(t *testing.T) ([]*Cluster, []*test.IpfsMock) {
 	clusters[0] = createCluster(t, hosts[0], dhts[0], cfgs[0], stores[0], cons[0], apis[0], ipfss[0], trackers[0], mons[0], allocs[0], infs[0], tracers[0])
 	<-clusters[0].Ready()
 	bootstrapAddr := clusterAddr(clusters[0])
+
 	// Start the rest and join
 	for i := 1; i < nClusters; i++ {
 		clusters[i] = createCluster(t, hosts[i], dhts[i], cfgs[i], stores[i], cons[i], apis[i], ipfss[i], trackers[i], mons[i], allocs[i], infs[i], tracers[i])
@@ -481,9 +488,7 @@ func pinDelay() {
 }
 
 func ttlDelay() {
-	diskInfCfg := &disk.Config{}
-	diskInfCfg.LoadJSON(testingDiskInfCfg)
-	time.Sleep(diskInfCfg.MetricTTL * 3)
+	time.Sleep(ttlDelayTime)
 }
 
 // Like waitForLeader but letting metrics expire before waiting, and

peer_manager_test.go

@@ -111,7 +111,7 @@ func TestClustersPeerAdd(t *testing.T) {
 	if err != nil {
 		t.Fatal(err)
 	}
-	pinDelay()
+	ttlDelay()
 
 	f := func(t *testing.T, c *Cluster) {
 		ids := c.Peers(ctx)
@@ -420,7 +420,7 @@ func TestClustersPeerRemoveReallocsPins(t *testing.T) {
 	// We choose to remove the leader, to make things even more interesting
 	chosenID, err := clusters[0].consensus.Leader(ctx)
 	if err != nil {
-		// choose a random peer
+		// choose a random peer - crdt
 		i := rand.Intn(nClusters)
 		chosenID = clusters[i].host.ID()
 	}
@@ -527,7 +527,7 @@ func TestClustersPeerJoin(t *testing.T) {
 
 	h := test.Cid1
 	clusters[0].Pin(ctx, h, api.PinOptions{})
-	pinDelay()
+	ttlDelay()
 
 	for _, p := range clusters {
 		t.Log(p.id.String())
@@ -574,7 +574,7 @@ func TestClustersPeerJoinAllAtOnce(t *testing.T) {
 
 	h := test.Cid1
 	clusters[0].Pin(ctx, h, api.PinOptions{})
-	pinDelay()
+	ttlDelay()
 
 	f2 := func(t *testing.T, c *Cluster) {
 		peers := c.Peers(ctx)